Qwen 3.6 Plus (High): benchmark results
Provider: Alibaba. Released 2026-04-02. Access: API.
Unified ELO 1675 ± 32, rank #430 of 2133 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| InfiniteBM Heads-Up No-Limit Hold'em | 1330.72 | Arena Elo (elo) | 81.8 |
| LivingScreen | 44.9 | Success rate (%), mean of the three task tiers; LivingScreen | 50 |
| LivingScreen - Closed-Loop Application | 22.8 | Success rate (%) on L3 fact-checking, content moderation and | 50 |
| LivingScreen - GUI Action | 72.8 | Success rate (%) on L1 atomic GUI actions (liking, collectin | 50 |
| BaFCo - Coarse Layout Analysis (CoT) | 6.93 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 44.4 |
| BaFCo - Coarse Layout Analysis (Zero-shot) | 5.89 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 44.4 |
| BaFCo - Layout Analysis (CoT) | 1.66 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 44.4 |
| BaFCo - Layout Analysis (Zero-shot) | 1.75 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 44.4 |
| BackendForge (Base Oracle) | 23.21 | Task success rate (%; base oracle of 7,250 items written bef | 42.3 |
| The Aggregate - Long-tail Book Recall | 31.9 | Accuracy (%) | 29.4 |
| LivingScreen - Cross-Source Understanding | 39.1 | Success rate (%) on L2 multiple-choice questions linking vid | 25 |
| The Aggregate - Exotic Fact Recall | 31.8 | Accuracy (%) | 18.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-6-plus-high · How It Works · Data refreshed daily, snapshot 2026-10-11.